You've pretty much just repeated what I already said. If you have a breakeven system, using a fixed bet size gets you nowhere on average and using a fractional bet size actually loses you ground in the long run. So the system has to be tweaked or replaced to get a new system with positive expectation for a fixed bet size, then you can apply Kelly sizing for optimal geometric growth, or some other sizing strategy if you have another utility function in mind.
I believe in the original thread IronFist was using 5 minute charts. I would imagine 1 hour or smaller (but not too small) would work just fine for this. What do you mean by "drift"? What does Kirshenbaum mean? The alphabet?
I remember reading about a system on a Forex forum years ago called "Lucky 7" or something like that where he basically had a trend following system and would increase the position size each time he opened a trade until he had a profitable trade. I think eventually he revamped he system to go up to 21 attempts or something. Not that I would do that, but what you said reminded me of that. If you could demonstrate conclusively that you won't have more than x losses, this would work. If price is random, then eventually you would blow your account. Here's an illustration from one of the threads: The right type of choppy environment could run through the trades really quick, though. Same thing if your broker widened their spread (since the original context of this was forex).
Here's an illustration of when this type of system fails. This isn't a 60 period MA like IronFist was talking about, but this type of thing happens with MAs of all lengths from time to time. The circled area is where you get chopped to death and give back all your gains and then some.
I coded this fast this afternoon - hopefully no system gotchas in the code, that are skewing results. EURO FX April'12 to present: $110K profit on 5 contracts (76K for '12 and 33K for '13 so far). I flipped through a range of MAs: The best MA level was using the 800 ema. NQ/ES didn't come close to those results, YM better. This was super fast down and dirty: 5 min 24 hr continuous contract, with filters I typically switch on for my own strats: Exit on Friday or Exit after 3:45pm EST, or Exit on trade equity > 1.0. Strong believer in time of day exits and day of week exits and didn't realize had these filters/functions in "on mode" as I borrowed the code from an existing strat meant for US hrs, not EUR. However, after removing those filters the results were dismal. max dd -21K (Feb 7 '12) W/L 4.6 P/F 1.5 (too low for me) (however, beautiful equity curve.) Threw in 1/2011 to present: market characteristics change so even though the W/L remains high, have PF dropping to 1.1, during that period you are gross P/L +685K and -605K with a crazy EQ curve. Gotta know when to break this baby out and when to put it in the dummy corner.
You can try to use it on 3 different time scale charts, trading the middle one when all 3 agree on the direction.
That would be a typo, it represents the point value currently open on the trade. Values started to even out after .061 on up to .1, that statement should read Exit Trade if Trade Equity for current trade > .1